Characterization of Neurospora crassa Tom40-deficient mutants and effect of specific mutations on Tom40 assembly.

The TOM complex (Translocase of the Outer mitochondrial Membrane) is responsible for the recognition of mitochondrial preproteins synthesized in the cytosol and for their translocation across or into the outer mitochondrial membrane. Tom40 is the major component of the TOM complex and forms the translocation pore. We have created a tom40 mutant of Neurospora crassa and have demonstrated that the gene is essential for the viability of the organism. Mitochondria with reduced levels of Tom40 were deficient for import of mitochondrial preproteins and contained reduced levels of the TOM complex components Tom22 and Tom6, suggesting that the import and/or stability of these proteins is dependent on the presence of Tom40. Mutant Tom40 preproteins were analyzed for their ability to be assembled into the TOM complex. In vitro import assays revealed that conserved regions near the N terminus (residues 51-60) and the C terminus (residues 321-323) of the 349-amino acid protein were required for assembly beyond a 250-kDa intermediate form. Mutant strains expressing Tom40 with residues 51-60 deleted were viable but exhibited growth defects. Slow growing mutants expressing Tom40, where residues 321-323 were changed to Ala residues, were isolated but showed TOM complex defects, whereas strains in which residues 321-323 were deleted could not be isolated. Analysis of the assembly of mutant Tom40 precursors in vitro supported a previous model in which Tom40 precursors progress from the 250-kDa intermediate to a 100-kDa form and then assemble into the 400-kDa TOM complex. Surprisingly, when wild type mitochondria containing Tom40 precursors arrested at the 250-kDa intermediate were treated with sodium carbonate, further assembly of intermediates into the TOM complex occurred, suggesting that disruption of protein-protein interactions may facilitate assembly. Import of wild type Tom40 precursor into mitochondria containing a mutant Tom40 lacking residues 40-48 revealed an alternate assembly pathway and demonstrated that the N-terminal region of pre-existing Tom40 molecules in the TOM complex plays a role in the assembly of incoming Tom40 molecules.

Most mitochondrial proteins are nuclear gene products that must be synthesized on cytosolic ribosomes, imported into mitochondria, and sorted to the correct mitochondrial subcompartment. These processes require the concerted action of complex protein translocases located in the outer and inner mitochondrial membranes (1)(2)(3). The process of importing mitochondrial preproteins into the organelle is initiated by the TOM complex (Translocase of the Outer mitochondrial Membrane), 1 which recognizes mitochondrial preproteins in the cytosol. The TOM complex facilitates the direct insertion of preproteins targeted to the outer membrane as well as the passage of preproteins destined for the inner mitochondrial compartments across the outer membrane. Further import and sorting of these preproteins requires the action of the TIM complexes (Translocases of the Inner mitochondrial Membrane).
Tom40 has been shown to be an essential protein in S. cerevisiae (16). The protein can be cross-linked to precursor proteins as they pass through the translocation pore (17,18) and is the major component of the TOM complex pore in both S. cerevisiae and N. crassa (4,19,20). Based on cross-linking studies to presequences, Tom40 may also form the major portion of the precursor binding trans site on the intermembrane space side of the outer membrane and contribute to the cis binding site on the cytosolic side of the membrane (21). Purified TOM complex and pores containing only Tom40 have similar structural and electrophysiological properties (19,22). The TOM complex pore has been found to be about 20 -26 Å in diameter by both electron microscopic analysis and size exclusion studies (4,5,19,22,23).
In the TOM complex, Tom40 exists as an oligomer with dimers as the basic structure (5,8,24,25). However, crosslinking studies have shown that, as precursors translocate, both the oligomer and the dimer undergo changes that affect the spatial interactions between different Tom40 molecules and between Tom40 and other TOM complex components (24). Predictions of Tom40 structure have suggested that the N and C termini extend into the intermembrane space. This is consistent with experimental evidence showing that both termini can be removed by added protease but only when the mitochondrial membrane is opened to allow the protease access to the intermembrane space (20). The remainder of the structure has been predicted to exist as a ␤-barrel, similar to bacterial porins (26,27), with 14 anti-parallel ␤-strands spanning the mitochondrial outer membrane. A high level of ␤-sheet was seen in Tom40 expressed in bacteria and refolded from exclusion bodies (19), but spectral analysis of Tom40 purified directly from mitochondria revealed less ␤-sheet and more ␣-helix than predicted (22).
The integration and assembly of Tom40 itself into the mitochondrial outer membrane requires the TOM complex and is only accomplished efficiently if the protein exists in a partially folded state (29). Assembly of Tom40 into the TOM complex is thought to occur via translocation intermediates (29 -31). In the first step of translocation, the Tom40 precursor binds at the outer surface of the TOM complex as a monomer. This monomer is imported through the outer membrane and assembled on the intermembrane space side of the membrane into an intermediate of 250 kDa that also contains pre-existing molecules of Tom40 and Tom5. There are two views for the mechanism by which Tom40 integrates into the outer membrane. In one model, integration occurs when the precursor that is associated with other components in the 250-kDa intermediate progresses to a 100-kDa intermediate that most likely contains Tom40 as a dimer of the newly imported subunit and a preexisting molecule. This intermediate undergoes further assembly and becomes associated with other Tom proteins to give the fully assembled TOM complex (30). In a second model, Tom40 was thought to insert into the membrane directly from the 250-kDa form into the 400-kDa fully assembled complex (29). A conserved set of amino acid residues near the N terminus of Tom40 is required for assembly and stability of Tom40 within the TOM complex, but is not involved in targeting of newly synthesized Tom40 to mitochondria (32).
To date, there has been only one report of cells lacking Tom40. S. cerevisiae cells depleted of the protein (then named Isp42p) were shown to accumulate mitochondrial precursors in the cytosol and ascospores lacking a functional copy of the gene were inviable (16). To examine and more fully characterize the effects of Tom40 depletion in another organism, we have generated a null allele of the gene in N. crassa that is maintained in a sheltered heterokaryon. We have also further investigated the assembly pathway of Tom40 into the TOM complex.

Growth of N. crassa and Strains
Used-Growth and handling of N. crassa strains was carried out as described previously (28). Strains used in this study are listed below in Table I. Race tubes were constructed as described previously (33).
Creation of Sheltered RIP Mutants-Repeat induced point mutation (RIP) occurs in N. crassa when one of the nuclei involved in a sexual cross carries a duplicated DNA sequence. RIP results in the generation of GC to AT transitions in both copies of the duplication (34). Because Tom40 was predicted to be an essential protein in N. crassa, the procedure of sheltered RIP was used to mutate the tom40 gene. Sheltered RIP insures that nuclei containing non-functioning alleles are present in a heterokaryon that also contains nuclei with a wild type copy of the gene. The rationale and strains utilized for sheltered RIP are described in detail elsewhere (35,36). Briefly, molecular restriction fragment length polymorphism mapping studies (37, 38) using a tom40-containing cosmid as a probe revealed that the tom40 gene was located on linkage group V of N. crassa. To create a duplication of tom40, spheroplasts of the HostV strain for sheltered RIP were transformed with a plasmid (pRIP-4) carrying hygromycin resistance (39) and 1.8 kb of PCR-generated tom40 genomic sequence. Transformants were selected on hygromycin, and strains containing duplications were identified by Southern analysis. One strain, 40Dupl, was crossed to the MateV strain. Any ascospores from this cross that were capable of growth on minimal medium were determined to be heterokaryons. One nucleus of the heterokaryon was wild type with respect to Tom40 function, whereas the other could contain either RIPed or wild type tom40 alleles. The desired heterokaryotic strain is shown in Fig. 1.
Several ascospore isolates from the sheltered RIP cross were screened for the predicted characteristics of tom40 RIP -sheltered heterokaryons. When heterokaryons with the genotype shown in Fig. 1 were grown in media containing lysine, leucine, and cycloheximide (concentrations ranging from 30 to 50 g/ml) the RIPed nucleus was forced to predominate the heterokaryon to provide resistance to the antibiotic. As a result, the growth rate was severely reduced, because Tom40 could not be supplied by the nucleus providing cycloheximide resistance. One strain (RIP40het), which grew slowly under these conditions, was analyzed further. Southern analysis (not shown) indicated that, in addition to the gene at the endogenous tom40 locus, the ectopically integrated RIP substrate was also present in the strain. Using specific primers, both RIPed versions of the gene were amplified by PCR. The gene at the endogenous locus was cloned and sequenced entirely. A total of 92 RIP mutations, including a RIP generated stop codon at residue 35, were identified. About 270 base pairs of the ectopic tom40 sequence were determined directly from specific PCR products, and 9 RIP mutations were observed, including one that created a stop codon at residue 30. For both RIPed alleles, the stop codons occurred prior to the first predicted membrane-spanning domain (26,27) of the protein, and we consider these to be effectively null alleles. We have briefly described the use of this strain for developing mutants expressing exclusively variants of Tom40 in a previous report (32).
Transformation of N. crassa-DNA was transformed into N. crassa using spheroplasts as described previously (40,41) or by electroporation of conidia using modifications of a previously described technique (42,43). For electroporation, conidia (1 week old) were harvested, washed three times with 1 M sorbitol, and resuspended in 1 M sorbitol at a concentration of 2 to 2.5 ϫ 10 9 conidia/ml. Linearized plasmid DNA (5 g in a final volume of 5 l) was mixed with 40 l of conidia, placed in a pre-chilled electroporation cuvette, and incubated for 5 min on ice. A Gene Pulser (Bio-Rad, Hercules, CA) was used with settings of 2.1 kV, 475 ⍀, 25 microfarads. Immediately after the pulse (time constant, 11-12 ms), 1 ml of ice-cold 1 M sorbitol was added and the conidia were allowed to recover for 1 h at 30°C. Aliquots (10 -100 l) of the mixture were added to top agar containing appropriate antibiotics for selection of transformants, and the mixture was poured onto plates containing the same medium.
Creation of Strains Expressing Mutant Variants of Tom40 -The method for development of strains expressing only mutant versions of Tom40 was described previously (32). Briefly, mutant alleles of tom40 were constructed by site-directed mutagenesis of single-stranded DNA derived from a Bluescript plasmid derivative containing a genomic version of N. crassa tom40 and a bleomycin resistance gene (44). Plasmids confirmed to carry the desired mutations were transformed into the tom40 RIP -sheltered heterokaryon (RIP40het). Plasmids used to develop strains expressing altered Tom40s in this study contained deletions in the region between residues 50 and 60 of the N. crassa Tom40 protein. Plasmids pRD, pRDTLL, and p297 lack the residues, RD, RDTLL, or all of residues 51-60, respectively. Another plasmid, p16.6, encoded a Tom40 variant where residues KLG at position 321-323 were changed to alanine residues. Transformants were selected on media containing cycloheximide (50 g/ml), lysine, and leucine (to select for transformants of the nucleus carrying the tom40 RIP alleles), as well as bleomycin (1.5 g/ml) and caffeine (0.5 mg/ml, to enhance the action of bleomycin), purified through one round of single-colony isolation on the same selective medium and tested for nutritional requirements. Plasmids giving rise to homokaryons requiring lysine and leucine contain mutant alleles capable of restoring Tom40 function to a level sufficient for viability. The presence of mutant alleles in the transformants was confirmed by sequencing tom40-specific PCR products generated from genomic DNA.
Creation of Tom40 Variants for in Vitro Import Studies-A tom40 cDNA was cloned into plasmid pGEM7Zf(ϩ). Mutant alleles were created by site-directed mutagenesis and used for in vitro transcription and translation to produce mutant Tom40 precursor proteins for import into isolated mitochondria.
Import of Radiolabeled Proteins into Isolated Mitochondria-For in vitro import studies, the isolation of mitochondria (45), import of Tom40 precursors (32), and import of other mitochondrial precursor proteins (46) were as described. Import was analyzed by SDS-PAGE or blue native gel electrophoresis (BNGE) and gels were viewed by autoradiography.
For some experiments, carbonate extraction was performed to determine if imported precursor proteins were inserted into membranes. Mitochondria were suspended in 0.1 M sodium carbonate (pH 11) for 30 min at 0°C. The mixture was centrifuged at 20,000 rpm in a TLA55 rotor (Beckman Instruments, Palo Alto, CA) for 30 min at 2°C, and the pellets were processed for BNGE.
Electron Microscopy-One milliliter of cell suspension from liquid cultures was fixed in 1.5% KMnO 4 for 30 min at room temperature followed by several washes with distilled water. The sample was then suspended in 0.05 M sodium cacodylate buffer containing 2% glutaraldehyde and 15% sucrose. The cells were pelleted by brief centrifugation at room temperature in a clinical centrifuge and were resuspended in the same buffer. After a 30-min incubation on ice, cells were post-fixed in 1% (w/v) OsO 4 and 1.5% (w/v) K 2 Cr 2 O 7 for 90 min on ice. Samples were then post-stained in 1% (w/v) uranyl acetate overnight at room temperature. The steps of dehydration, embedding, and sectioning were performed by the Microscopy Unit, Department of Biological Sciences, University of Alberta. Sections were examined in a transmission electron microscope.
Other Techniques-The standard techniques of agarose gel electrophoresis, Southern and Northern blotting of agarose gels, preparation of radioactive probes, transformation of Escherichia coli, isolation of bacterial plasmid DNA, and the PCR using a mixture of Taq and Vent polymerase (New England BioLabs, Beverly, MA) to minimize replication errors, were all performed as described (49). The following procedures were employed using the supplier's recommendations or previously described procedures: isolation of total RNA with the Qiagen RNeasy plant mini kit (Qiagen Inc., Santa Clarita, CA), separation of mitochondrial proteins by polyacrylamide gel electrophoresis (50), Western blotting (51), Western blot detection using LumiGLO chemiluminescent substrate (Kirkegaard and Perry Laboratories, Gaithersburg, MD), genomic DNA extraction (52), protein determination with the Coomassie dye binding assay (Bio-Rad, Hercules, CA), manual DNA sequencing using thermosequenase (Amersham Biosciences, Cleveland, OH), and automated sequencing using a DyeNamic sequencing kit (Amersham Biosciences) with a Model 373 stretch sequencer separation system (Applied Biosystems, Foster City, CA), and site-directed mutagenesis using the Muta-Gene system (Bio-Rad). Radioactive precursor proteins for import were generated by coupled in vitro transcription and translation with the Promega (Madison, WI) TNT reticulocyte lysate system in the presence of [ 35 S]methionine (ICN, Costa Mesa, CA).

Isolation of N. crassa tom40 Null Mutants-Because
Tom40 is the major component of the mitochondrial outer membrane translocation pore (22) and is an essential gene in yeast (16), we used the procedure of sheltered RIP to obtain N. crassa tom40 null mutants. The product of the procedure was a heterokaryotic strain (designated RIP40het) in which the tom40 gene in one nucleus is inactivated by RIP, whereas the other nucleus retains a wild type copy of the gene (Fig. 1). Sequence analysis revealed that the tom40 RIP -bearing nucleus contains only null alleles of the gene (see "Materials and Methods"). To determine if tom40 is an essential gene in N. crassa, conidiaspores produced by RIP40het were streaked onto medium containing all the nutritional requirements of both nuclei in the heterokaryon ( Fig. 1 and Table I). N. crassa conidiaspores are usually multinucleate and the heterokaryon should produce three separate types of conidiaspores, assuming random segregation of nuclei into conidia: homokaryons for the lysine-leucine-requiring tom40 RIP nucleus, homokaryons for the inositol-requiring sheltering nucleus, and heterokaryons containing both nuclei. If tom40 is essential, no lysine-leucine-requiring conidia should be viable. Testing of nutritional requirements of 181 individual colonies isolated from these plates revealed that 120 were heterokaryons, 61 were inositol-requiring homokaryons, and none were lysine-leucinerequiring homokaryons. Thus, the tom40 RIP nucleus was inviable. To confirm that the effects of RIP were specific to the tom40 gene, the sheltered heterokaryon was transformed with a bleomycin resistance plasmid containing a wild type copy of tom40. When selected on media containing lysine, leucine, cycloheximide, and bleomycin, viable lysine-leucine-requiring homokaryotic strains were recovered. Considered together, these data show that tom40 is an essential gene in N. crassa.
Characteristics of tom40-deficient Cells-The two nuclei in RIP40het differ with respect to auxotrophic and antibiotic resistance markers, which make it possible to force the tom40 RIP nucleus to predominate the heterokaryon by growth in medium FIG. 1. Sheltered heterokaryon for the tom40 RIP mutant. The box symbolizes the heterokaryotic RIP40het strain used in this study, containing two distinct nuclei whose genotypes are enclosed by circles. Nucleus 1 contains no functional copies of tom40; nucleus 2 contains a wild type allele of the gene. Other markers necessary for the manipulation of the heterokaryon are indicated. The mutant allele of the cyh-2 gene provides resistance to cycloheximide. The heterokaryon was produced as described under "Materials and Methods." As HV, but carries an ectopic copy of tom40; hygromycinresistant Transformation of HostV with the tom40 containing plasmid pRIP-4 RIP40het Sheltered heterokaryon: (cyh-2 lys-2 leu-5 mei-2 tom40 RIP ϩ am132 inl inv mei-2), mating type unknown. Both nuclei also contain an ectopic RIPed copy of tom40.
Transformation of RIP40het with plasmid p297 321AAA cyh-2 lys-2 leu-5 mei-2 tom40 RIP tom40 RIP (EC), also contains an ectopic copy of tom40 with residues KLG at positions 321-323 changed to alanine residues Transformation of RIP40het with plasmid p16.6 containing lysine, leucine, and cycloheximide ( Fig. 1). Under these conditions, the growth rate of RIP40het was clearly reduced ( Fig. 2A), because the cycloheximide resistance-conferring tom40 RIP nucleus was unable to supply sufficient levels of Tom40 when it predominated in the heterokaryon. Analysis of mitochondrial proteins in RIP40het grown under these conditions showed the predicted reduction in Tom40 (Fig. 2B). BNGE of the TOM complex from Tom40-deficient mitochondria revealed no differences in size or stability (Fig. 2C), demonstrating that the complex forms normally in the mutant but is simply reduced in amount. Mitochondria deficient in Tom40 were also examined with respect to the levels of other mitochondrial proteins, including other members of the TOM complex. The level of the TOM holo-complex receptor proteins Tom70 and Tom20, as well as the mitochondrial proteins Hsp70 and porin were unaffected by the reduction in Tom40 levels. However, the amounts of the TOM core complex proteins, Tom22 and Tom6, were severely reduced (Fig. 2B). To rule out the possibility that the Tom40 deficiency might signal a down-regulation of transcription of the tom22 and tom6 genes, we analyzed Northern blots for the presence of tom40, tom22, and tom6 mRNAs. The level of tom40 transcript is severely reduced in RIP40het cells grown under conditions where the tom40 RIP nucleus predominates the culture (Fig. 3). This was expected, because genes that have undergone severe RIP produce little or no transcript (53,54). We were unable to detect a transcript for tom6 in any control or mutant strain (not shown), suggesting that few transcripts are produced and/or that they are short-lived. However, transcripts of tom22 were present in normal amounts in Tom40-deficient cells (Fig. 3) so that the decreased levels of this protein, and by analogy Tom6, were not due to decreased levels of transcription. Thus, the data suggest that either the import and/or stability of Tom22 and Tom6 in the membrane is/are dependent on assembly into the core complex with Tom40. It should be noted that the reduction in these two components was not due to a generalized decrease in import capacity that would be expected in Tom40-deficient mitochondria, because other mitochondrial proteins were present at normal levels (Fig. 2B). It is likely that the reduced growth rate of the cells reflects the rate at which mitochondrial proteins can be imported so that their steady-state levels do not differ significantly from wild type cells. This interpretation is supported by the observation that import of mitochondrial precursors into isolated Tom40-deficient mitochondria was reduced (Fig. 4).
To determine the structure of Tom40-deficient mitochondria, cells depleted of the protein by growth in cycloheximide were examined by electron microscopy. Mitochondria with reduced levels of Tom40 were smaller than controls and contained virtually no cristae (Fig. 5).

FIG. 2. Characteristics of Tom40-deficient cells.
A, the control strain (40Dupl) was grown in the presence of lysine and leucine (control, diamonds) and RIP40het was grown in minimal medium (RIP40het, triangles) to compare their growth rates under conditions where complementation of Tom40 function occurs in the sheltered heterokaryon. To assess the growth rate of Tom40-deficient cells, the RIP40het strain was grown in medium containing lysine, leucine, and cycloheximide at 50 g/ml (RIP40het ϩ CHI, squares), which forces the tom40 RIP nucleus to predominate the culture. The cycloheximide resistant control strain was grown under similar conditions (control ϩ CHI, crosses). B, mitochondria were isolated from the indicated strains grown under the conditions described in A. Mitochondrial proteins were separated by SDS-PAGE, blotted to nitrocellulose, and immunodecorated with antisera directed against the indicated proteins (mt, mitochondrial). C, mitochondria from the control strain (40Dupl) and RIP40het grown in the presence of lysine, leucine, and cycloheximide were solubilized in either digitonin (DIG) or dodecyl maltoside (DDM) and analyzed by BNGE. The gel was blotted to PVDF membrane and immunodecorated with antiserum against Tom40. The positions of molecular weight markers are indicated on the left.

A Region Near the N Terminus of Tom40 Affects Assembly and Stability of the TOM Complex-
Little is known about the domains and amino acid residues of Tom40 that are necessary for it to accomplish its functions or for its own assembly into the TOM complex. Alignment of the protein from several organisms reveals the existence of several conserved residues (Fig.  6). We had previously shown that a conserved region near the N terminus (residues 40 -48) was important for assembly of N. crassa Tom40 into the TOM complex (32). The region is followed by another set of residues (51-60) that are very well conserved between yeast and N. crassa and reasonably well conserved between fungi and animals (Fig. 6). To determine if these residues were also important for the assembly of Tom40, we developed cDNA versions of the gene-encoding variants lacking the RD residues (⌬RD), the RDTLL residues (⌬RDTLL), and the ten residues from position 51-60 (⌬51-60) (Fig. 7A). These were transcribed and translated in vitro to generate Tom40 precursor proteins for import into isolated mitochondria. Wild type Tom40 precursor has been shown to accumulate in a 250-kDa intermediate when imported into isolated mitochondria at 0°C (29,32). The most recent model of Tom40 assembly suggests that the precursor further assembles into the 400-kDa TOM complex via a 100-kDa intermediate during import at 25°C (30). All three deletion variants had a reduced capacity to be imported into the fully assembled TOM complex at 25°C and tended to accumulate in the high molecular mass 250-kDa assembly intermediate (Fig. 7B). A fraction of the ⌬RD and ⌬RDTLL precursors did reach the assembled state, suggesting a relatively mild affect on assembly. However, the ⌬51-60 variant was more severely affected, and none of this variant became fully assembled after 20 min of import at 25°C. There was a similar gradient of effect on the formation of the 100-kDa intermediate, which was not present as a discrete band at either temperature when the ⌬51-60 variant was imported.
Genomic versions of all three mutant variants were found to rescue the tom40 RIP nucleus of strain RIP40het. Strains containing the ⌬RD and ⌬RDTLL forms of Tom40 exhibit only slightly reduced growth rates and an inability to climb the walls of growth flasks (not shown). Despite the finding that virtually none of the ⌬51-60 variant assembled into the TOM complex in vitro, the rate of assembly in vivo must be sufficient to allow viability, because strains expressing only this version of Tom40 were obtained. However, growth defects in the ⌬51-60 strains were evident and were enhanced at 15°C where the deletion strains grew at roughly half the rate of controls (Fig. 7C). The ⌬51-60 strains also produced fewer conidiaspores than the strains containing the milder variants (not shown). Thus, the level of assembly of these Tom40 variants that was observed in vitro correlates with the severity of phenotype observed in vivo. In addition, when analyzed by BNGE, the TOM complex in each of the deletion strains appears to be more fragile than in controls when mitochondria are solubilized with dodecyl maltoside and break down into a series of smaller complexes (Fig. 7D). A striking deficiency of Tom40 also appears in the lanes containing mitochondria from the ⌬51-60 strain. When mitochondria from this strain were solubilized in SDS, no difference in the amount of Tom40 was observed relative to controls (Fig. 7E). Therefore, the deficiency of Tom40 observed following BNGE may have been due to aggregation of the complex from this strain, resulting in an inability to enter the gel.
A Conserved C-terminal Region of Tom40 Is Also Required for Assembly-To examine the effects of mutations outside the N-terminal region on Tom40 assembly, we chose a block of residues in the most C-terminal region of similarity between all the species shown in Fig. 6. Mutant versions of N. crassa tom 40 cDNA in which the residues KLG at position 321-323 were either deleted (⌬321-3) or changed to Ala residues (321AAA) were constructed (Fig. 8A). The variants were transcribed and translated in vitro to produce mutant precursor proteins for use in import studies. The import and assembly of the 321AAA precursor was similar to the wild type precursor, although assembly to the final 400-kDa complex was somewhat less efficient (Fig. 8B). On the other hand, the ⌬321-3 variant did not progress to the 250-kDa intermediate during import at 0°C and did not reach the fully assembled complex at either temperature. Thus, the deletion of the KLG residues dramatically affects the ability of Tom40 precursors to assemble into the TOM complex, whereas changing the residues to alanines has little effect. In keeping with the severe assembly defects observed in vitro, we were unable to rescue the tom40 RIP nucleus with constructs containing the ⌬321-3 construct. However, using the 321AAA construct, strains with growth defects were isolated (Fig. 8C). BNGE analysis of mitochondria isolated FIG. 4. Tom40-deficient mitochondria have reduced ability to import mitochondrial precursor proteins. Mitochondria were isolated from the control strain (40Dupl) and RIP40het following growth in lysine, leucine, and cycloheximide (30 g/ml) so that mitochondria in RIP40het contained reduced levels of Tom40. Import of radiolabeled precursors (F 1 ␤, the ␤-subunit of the F 1 -ATPase; MPP, the mitochondrial processing peptidase; and AAC, the ATP/ADP carrier protein) was performed at 20°C for the times shown. Following a post-import treatment with proteinase K to remove non-imported precursors, mitochondria were re-isolated and subjected to SDS-PAGE. The gels were blotted to nitrocellulose and exposed to x-ray film. The lysate lane contained 33% of the input lysate used in each import reaction. The precursor (p) and mature (m) forms of MPP and F 1 ␤ are indicated. One sample from each strain was treated with trypsin prior to import (Pre-trypsin) to demonstrate that import was receptor-dependent.
FIG. 5. Appearance of Tom40-deficient mitochondria. The cycloheximide-resistant control strain (40Dupl) and RIP40het were grown under the conditions described in the legend to Fig. 2A. At the indicated times, mycelium was harvested and processed for electron microscopy as described under "Materials and Methods." from 321AAA strains revealed striking defects in TOM complex stability (Fig. 8D). When solubilized in the mild detergent digitonin, a substantial fraction of Tom40 was not present in the 400-kDa fully assembled form and migrated at a position that would correspond roughly to a Tom40 dimer. From these data it is impossible to determine if the smaller form of the complex forms during solubilization and BNGE or if it exists in vivo. Solubilization of mitochondria containing the 321AAA Tom40 variant in dodecyl maltoside results in Tom40 being released as a monomer. We have previously shown that the C-terminal extension in fungal Tom40 proteins (residues 330 -349 of the N. crassa protein, Fig. 6) is not required for assembly (32) or function of the TOM complex. 3 TOM complex containing the variant lacking this fungal C-terminal extension (residues 330 -349, Fig. 6) was virtually indistinguishable from controls in terms of TOM complex stability on blue native gels (Fig. 8D). Thus, instability of the complex is not caused by general perturbations of the protein at the C terminus. The role of residues 321-323 in the function and stability of the TOM complex was further demonstrated by the fact that mitochondria containing the 321AAA variant imported mitochondrial precursors less efficiently than control mitochondria (Fig. 8E).
Nature of TOM Complex Assembly Intermediates-We consistently observed a high molecular weight band on our blue native gels of Tom40 import experiments that had not been discussed in previous reports. We estimated the band to be ϳ500 kDa in size. The band did not appear in substantial amounts when import was stalled at the 250-kDa stage ( Fig.  7B and 8B). To assess the relevance of the 500-kDa band we performed import experiments where Tom40 precursor was added for only a 4-min pulse at 25°C. The kinetics of precursor assembly were then followed over a period of 240 min at 25°C. The experiment showed that the amount of the 500-kDa form did not change appreciably over the course of the experiment (Fig. 9). Thus, the 500-kDa band most likely represents Tom40 precursor in a non-productive state. This experiment also showed the precursor-to-product relationship between the 100and 250-kDa intermediates and the fully assembled 400-kDa form, because the amount of radioactive Tom40 precursor in the intermediates gradually decreases as more radioactivity accumulates in the 400-kDa form (Fig. 9). When the levels of radioactivity in the intermediates were quantified for each time point, we could not demonstrate that the appearance of Tom40 precursor in the 250-kDa form precedes its appearance in the 100-kDa intermediate as previously reported (30). In our experiments the Tom40 precursor appeared in both intermediates simultaneously. The rate of disappearance of the precursor from each intermediate, as Tom40 assembled to the 400-kDa form, was also virtually identical. Attempts to delay the appearance of the intermediates by lowering the temperature of import to 10 or 15°C, and shortening the pulse time to 1 min, 3 R. D. Taylor and F. E. Nargang, unpublished observations. reduced the rate of assembly but revealed no differences in precursor product relationships (not shown). However, results obtained during studies of import and assembly of mutant Tom40 precursors did support the hypothesis that the 250-kDa intermediate is formed prior to the 100-kDa form (see below).
Carbonate Treatment Enhances the Assembly of Wild Type Tom40 Precursor into the TOM Complex-Based on the results of sodium carbonate extraction experiments it was previously suggested that a newly imported Tom40 precursor in the 250-kDa intermediate is peripherally associated with the mitochondrial outer membrane in the intermembrane space. On the other hand, when the precursor progresses to the 100-kDa intermediate and the 400-kDa assembled form, it becomes an integral membrane protein (30). We attempted to confirm the carbonate extractability of Tom40 precursor in the 250-kDa intermediate and did observe that the amount of the intermediate form was reduced following extraction of imports performed at 25°C. However, the interpretation of these data was not straightforward, because we also observed that newly imported Tom40, which accumulates at intermediate steps when import is performed at 0°C, was assembled into the 400-kDa form of the TOM complex as a result of carbonate treatment. This is shown in Fig. 10 where the expected accumulation of Tom40 precursor in the 250-kDa form during import at 0°C was seen in the untreated samples and very little of the fully assembled 400-kDa complex was present, even after 60 min of import (Fig. 10A, Tom40wt lanes). This contrasts with results obtained when import was performed at 25°C, when most of the newly imported precursor was present in the fully assembled form (Fig 10A, Tom40wt lanes). However, after mitochondria were subjected to carbonate extraction, much of the precursor imported at 0°C was present in the fully assembled complex (Fig. 10B, Tom40wt lanes). The trivial explanation that further assembly to the 400-kDa form occurs as a result of the extra time at 0°C during incubation with sodium carbonate cannot account for the data, because the untreated samples (panel A) were handled in an identical fashion except for the  Fig. 6 is shown to illustrate the location and residues affected in the Tom40 variants ⌬RD, ⌬RDTLL, and ⌬51-60. Deleted residues are indicated by dashes. For the ⌬51-60 mutant, residue 50 was also changed from a Glu to an Ala (lowercase a). Organisms in the comparison are as defined in the legend to Fig. 6. B, radiolabeled precursors of a wild type Tom40 (Tom40wt) and the indicated variants were incubated at either 0°or 25°C with wild type (74A) mitochondria for 20 min. Mitochondria were re-isolated and solubilized in buffer containing 1% digitonin. The samples were electrophoresed on blue native gels, blotted to PVDF membrane, and analyzed by autoradiography. The positions of the 400-kDa TOM complex and the 100-and 250-kDa intermediates are indicated. M, Tom40 monomer. C, growth phenotype of ⌬51-60 strains. Conidia from a ⌬51-60 strain (open squares) and a control strain (40Dupl, closed circles) were inoculated into flasks containing liquid medium, grown with shaking at 15°C, and harvested at the indicated times. D, BNGE analysis of N-terminal deletion strains. Mitochondria were isolated from a control (40Dupl) and strains expressing the Tom40 variants ⌬RD, ⌬RDTLL, or ⌬51-60. The organelles were solubilized with either digitonin (DIG) or dodecyl maltoside (DDM) and examined by BNGE. The gel was blotted to PVDF membrane and decorated with antiserum to Tom40. The position of molecular weight markers is indicated on the left. E, SDS-PAGE analysis. Mitochondria from control strain (40Dupl) and a ⌬RDTLL mutant strain were dissolved with Laemmli gel cracking buffer and subjected to SDS-PAGE. The gel was blotted to nitrocellulose and immunodecorated with antiserum to Tom40 and Tom22.

FIG. 8. Assembly of the ⌬321-3 and 321AAA variants of Tom40.
A, a region of the alignment from Fig. 6 is shown to illustrate the location and residues affected in the Tom40 variants 321AAA and ⌬321-3. Deleted residues are indicated by dashes, and amino acid changes are shown as lowercase letters. Organisms in the comparison are defined in the legend to Fig. 6. B, radiolabeled precursors of wild type Tom40 (Tom40wt) and variants ⌬321-3 and 321AAA were imported into wild type (74A) mitochondria for 20 min at either 0°or 25°C. Mitochondria were re-isolated and solubilized in 1% digitonin. The samples were electrophoresed on blue native gels, blotted to PVDF membrane, and analyzed by autoradiography. The positions of the 400-kDa TOM complex and the 100-and 250-kDa intermediates are indicated. M, Tom40 monomer. C, growth of the 321AAA mutant. A control strain (40Dupl, square symbols) and a 321AAA mutant strain (circles) were inoculated in race tubes and incubated at either 22°C (filled symbols) or 15°C (open symbols). The extent of mycelial elongation was measured daily. D, BNGE analysis of C-terminal mutant strains. Mitochondria were isolated from a control (40Dupl) and strains expressing the Tom40 variants 321AAA or a C-terminal deletion of Tom40 lacking residues 330 -349 (⌬C-term). Mitochondria were solubilized with either digitonin (DIG) or dodecyl maltoside (DDM) and examined by BNGE. The gel was blotted to PVDF membrane and decorated with antiserum to Tom40. The position of molecular weight markers is indicated on the left. E, import of precursors into 321AAA mitochondria. Import into mitochondria isolated from a control strain (40Dupl), and a 321AAA strain was performed as described in the legend to Fig. 4 using the precursors of MPP, the mitochondrial processing peptidase, and Su-9 dihydrofolate reductase, a fusion of the N-terminal 69 amino acids of N. crassa ATPase subunit 9 to the coding sequence of mouse dihydrofolate reductase (57). addition of carbonate. These data demonstrate that assembly of precursors into the 400-kDa form is stimulated by the addition of carbonate. Thus, we cannot confirm the carbonate extractability of Tom40 precursor from the 250-kDa intermediate, because the possibility that Tom40 in the intermediate may have assembled into the 400-kDa form during the carbonate treatment cannot be excluded. It was possible that the newly imported Tom40 in the 400-kDa complex formed during carbonate treatment did not contain correctly assembled Tom40. To test this possibility we treated samples with proteinase K, which gives rise to characteristic 26-and 12-kDa cleavage products when the protein is correctly assembled (29,32). These fragments can be clearly seen in samples not treated with sodium carbonate (Fig. 10C). The carbonate-treated samples also show evidence of these bands, although both are reduced in amount and the 26-kDa fragment is present as a series of shortened fragments (Fig. 10C). This is most likely because Tom40 in the membrane sheets created by the action of carbonate is fully exposed to the added proteinase and undergoes more degradation than Tom40 in intact mitochondria. We conclude that the material in the 400-kDa form that arose by the action of carbonate represents correctly assembled Tom40.
Mutant Tom40 Precursors in the 250-kDa Assembly Intermediates Are Extractable by Carbonate-The precursor of the ⌬51-60 Tom40 variant accumulated at the 250-kDa intermediate stage during import at both 0°and 25°C (Fig. 7B). We wished to determine if the variant protein arrested at this stage would also assemble into the 400-kDa complex as the result of carbonate treatment. As shown in Fig. 10A (⌬51-60  lanes), the mutant form accumulates at the 250-kDa stage even after 60 min of import at 25°C. Unlike the wild type precursor at this stage, the variant was not efficiently converted to the assembled form in the presence of sodium carbonate (Fig. 10B), and a substantial amount of the protein in the intermediate was removed (compare panels A and B, ⌬51-60 lanes). Thus, a variant precursor that is inefficient at progressing past the 250-kDa stage is largely carbonate-extractable. A small fraction of the ⌬51-60 variant does remain in the 250-kDa intermediate following carbonate treatment suggesting that at least some of the newly imported ⌬51-60 Tom40 molecules at the 250-kDa stage are inserted into the membrane. Another small amount may be converted to the 100-kDa form (Fig. 9B, compare the smear in the ⌬51-60 lanes near 100 kDa in panel A to the discrete bands in panel B). The 100-kDa form appears to have a faster electrophoretic mobility when the variant Tom40 precursor forms part of this intermediate.
Further confirmation that Tom40 precursor in the 250-kDa FIG. 9. Kinetics of TOM complex assembly. Import of radiolabeled wild type Tom40 precursor was allowed to proceed into wild type (74A) mitochondria for 4 min at 25°C. Mitochondria were re-isolated at 4°C and resuspended in fresh import mix containing no additional Tom40 precursor. Import was allowed to continue at 25°C, and aliquots were removed at the times indicated. Mitochondria were re-isolated from the aliquots and processed for BNGE. The gel was blotted to PVDF membrane and analyzed by autoradiography. The sizes and position of bands are indicated on the left.
FIG. 10. Sodium carbonate treatment results in assembly of Tom40 precursors into the TOM complex. For A and B, radiolabeled precursors of wild type Tom40 and variant ⌬51-60, were imported into wild type (74A) mitochondria at 0°and 25°C for the times indicated. The sample was divided equally, and mitochondria were pelleted. One tube from each import was held on ice while the mitochondria in the other were resuspended in 0.1 M sodium carbonate. After 30 min on ice, the membrane fraction from the carbonate-treated samples and the untreated mitochondrial samples were pelleted at 2°C. Both pellets were suspended in 1% digitonin and processed for BNGE. A, BNGE of untreated samples; B, BNGE of carbonate-treated samples; C, import of radiolabeled wild type Tom40 precursor into wild type (74A) mitochondria was done for 20 min at either 0°or 25°C. For each temperature the experiments were performed in quadruplicate, and mitochondria were pelleted. In the first sample, the pellets were solubilized in 1% digitonin and prepared for BNGE. In the second sample, the mitochondria were treated with sodium carbonate. The resulting membrane fraction was solubilized in 1% digitonin and processed for BNGE. In the third sample, the pellets were suspended in import buffer, and treated with proteinase K (0.1 g/l) for 15 min. Mitochondria were re-isolated and processed for SDS-PAGE. In the fourth sample, mitochondria were subjected to sodium carbonate extraction, and the membrane fraction was pelleted, resuspended in import buffer, and treated with proteinase K as for the third sample. The membranes were re-isolated and processed for SDS-PAGE. Samples not treated with proteinase K and analyzed by BNGE are shown on the left. Samples treated with proteinase K and examined by SDS-PAGE are on the right. Gels were blotted to PVDF membrane (BNGE) or nitrocellulose (SDS-PAGE) and analyzed by autoradiography. For all panels, the positions and size of bands are indicated on the left. M, Tom40 monomer. form is extractable with carbonate was obtained in similar experiments using the precursor form of the ⌬321-3 Tom40. This variant is also entirely extractable from the 250-kDa intermediate with carbonate and is not converted to the assembled form (Fig. 11, A and B). Even long exposures of the blots in Fig. 11 to x-ray film or a phosphorimaging screen did not reveal any of the 250-kDa intermediate or the 400-kDa assembled form in carbonate-extracted samples. There is also no indication that this variant of Tom40 reaches the 100-kDa intermediate stage. It appears that the ⌬321-3 form of Tom40 is incapable of progressing past the carbonate-extractable 250-kDa intermediate stage of the assembly pathway. Thus, assembly mutants that accumulate at the 250-kDa intermediate stage provide good evidence that the Tom40 precursor has only a peripheral association with the membrane at this stage.
The Assembly Characteristics of Mutant Tom40 Variants Provide Evidence That Formation of the 250-kDa Intermediate Precedes Formation of the 100-kDa Intermediate-Our results on the assembly of the ⌬321-3 and ⌬51-60 mutant variants of Tom40 showed that these forms of the protein accumulate at the 250-kDa intermediate stage of the pathway. Although a smear in the region from Tom40 monomer up to the 100-kDa form is often seen during the assembly of these mutants (Figs. 7,8,10,and 11), no clear band corresponding to the 100-kDa form seen in the import of wild type precursor is observed. These observations support the notion that formation of the 250-kDa form precedes the formation of the 100-kDa intermediate.
An Alternative Assembly Pathway for Tom40 in Mitochondria Containing Mutant TOM Complex-It was of interest to determine how effectively Tom40 precursors would be imported/assembled into mitochondria with TOM complex containing only mutant versions of Tom40. Wild type Tom40 precursor proteins were imported into mitochondria isolated from a strain developed by rescue of the null nucleus of RIP40het with the ⌬40 -48 variant of Tom40 described previously (32). The TOM complex in these mitochondria contains only the mutant version of the protein. Surprisingly, wild type Tom40 precursors accumulated in a larger form of about 450 kDa when imported into mitochondria of the mutant at 0°C (Fig. 12A). This band was not identical to the unproductive band at 500 kDa that was discussed earlier, as shown by comparing the control and ⌬40 -48 lanes in Fig. 12A. A small amount of precursor did exist in the usual 250-kDa intermediate, but the 100-kDa form was not observed (Fig. 12A). When import was performed for 20 min at 25°C, assembly to the 400-kDa form was quite efficient, but the usual intermediates were almost entirely absent. To more fully address this finding, we examined conversion of the wild type precursor to the assembled form over a time course of 0 -120 min of import into the ⌬40 -48 variant mitochondria. A small amount of the 250-kDa intermediate was observed, but a discrete band for the 100-kDa intermediate was not evident at any time point (Fig. 12B). The conversion of precursor molecules in the 450-kDa intermediate to the 400-kDa form occurred in a pattern that is consistent with a precursor to product relationship (Fig. 12B). Carbonate extraction showed that Tom40 precursor was removed from the higher molecular weight intermediate, but not the assembled form (Fig. 12B). As with import into wild type mitochondria, at least some of the Tom40 precursor was converted into the assembled 400-kDa form as the result of carbonate treatment. Digestion with proteinase K demonstrated that Tom40 was not assembled in its final conformation in the high molecular weight intermediate that was the most abundant form present at 0 min but was in the correct conformation in the 400-kDa form at subsequent time points as judged by the formation of the 26-and 12-kDa fragments (Fig. 12C). These results indicate that the assembly pathway of Tom40 precursor was influenced by the altered N-terminal intermembrane space domain of Tom40 molecules in the TOM complex of the ⌬40 -48 mutant mitochondria.

DISCUSSION
Tom40 was previously shown to be essential for the viability of S. cerevisiae cells and reduced levels of Tom40 resulted in the accumulation of mitochondrial precursors in the cytosol (16). We have confirmed these findings by showing that tom40 is an essential gene in N. crassa and that mitochondria containing reduced levels of Tom40 are deficient in their capacity to import precursor proteins in vitro. We have extended the original findings by showing that mitochondria with lowered levels of Tom40 are smaller than normal and are devoid of cristae. It is probable that growth and accumulation of mitochondria are limited by the deficiency of Tom40 and that the reduction in growth rate of cells with depleted levels of the protein is related to the reduced capacity to accumulate essential factors in the organelle. Similar mitochondria are observed in slow growing cells deficient in other important components of the TOM complex such as Tom20 and Tom22 (12,46). Mitochondria deficient in Tom40 also have reduced levels of the TOM core complex components Tom22 and Tom6. These deficiencies are most likely due to the inability to import and/or assemble the proteins in the absence of Tom40. The possibility that reduced levels of Tom40 might signal reduced transcription of these genes was eliminated by demonstrating that the mRNA for Tom22 is present at normal levels.
We have previously shown that a well conserved region near the N terminus of Tom40 results in TOM complex assembly defects (32). The previously described mutant, lacking residues 40 -48 of the N. crassa Tom40 protein, was less efficient at assembling into the complex than wild type Tom40 precursors. Here we show that an adjacent region has similar effects. The ⌬51-60 mutant form of Tom40 is unable to progress past the 250-kDa intermediate stage to the fully assembled TOM complex even after 60 min of import at 25°C. The deletion of residues 51-60 does not totally abrogate assembly in vivo, because strains expressing only this form of the protein are viable, although they display growth defects and alterations in TOM complex stability. Thus, the region encompassed by both the ⌬40 -48 and ⌬51-60 mutations plays an important role for achieving the interactions necessary for progression past the 250-kDa intermediate on the assembly pathway.
The lack of assembly to any high molecular weight intermediate of the ⌬321-3 variant at 0°C was striking. At this temperature the altered protein may not be in an import-competent conformation. At 25°C, the precursor can be imported to the 250-kDa intermediate but cannot proceed to the membrane integration step of the assembly pathway. The severity of the assembly defects in vitro are supported by the observation that a tom40 gene encoding a ⌬321-3 variant cannot rescue the tom40 null nucleus. It has been shown that Tom40 must be in a partially folded state for effective integration into the membrane (29) and the ⌬321-3 variant may be unable to achieve the correct conformation for integration. However, it seems unlikely that there are gross overall changes in conformation at 25°C, because the protein assembles past the initial recognition stage and reaches the 250-kDa intermediate in vitro. The KLG residues might provide a specific signal for integration into the membrane or assembly with another TOM complex component, but this seems unlikely, because changing the residues to alanines has little effect on in vitro assembly and a tom40 gene encoding the 321AAA variant can restore viability, although the resulting strains have growth and TOM complex defects. Finally, it is conceivable that the KLG residues form part of a membrane-spanning domain. In this case, a reasonable explanation for the drastically different behavior of the deletion and substitution variants could be that replacement with alanine residues still allows the region to span the membrane, whereas loss of the residues does not allow membrane spanning and prevents integration. Such a membrane-spanning region might also be important for channel formation.
Given its position near the C terminus and the striking fragility of the TOM complex from the 321AAA strain, it is conceivable that the region could also be involved in maintaining interactions between TOM complex subunits. The presence of the Tom40 dimer-sized complex in digitonin-treated samples is reminiscent of the subcomplexes of Tom40 dimers seen in Tom22-deficient yeast cells (55). Thus, the 321AAA Tom40 mutant might have weakened interactions with Tom22. The breakdown of the complex in 321AAA mitochondria to Tom40 monomers following solubilization with dodecyl maltoside suggests that the region could also contribute to the formation and maintenance of Tom40 dimers.
It has been suggested that Tom40 precursor appears in the 250-kDa intermediate prior to the 100-kDa form on the assembly pathway (30). We were unable to confirm this in time course experiments using wild type Tom40 precursor, because both forms appeared and decreased concomitantly. When the assembly of mutant Tom40 molecules was examined, many variants gave rise to both the 250-and the 100-kDa intermediates. Because both intermediates can be seen during the import of wild type and various mutant forms of the precursor, it appears that neither represents a strictly rate-limiting step on the assembly pathway. Nonetheless, the existence of the 250-kDa form with no evidence of the 100-kDa intermediate in the ⌬51-60 and ⌬321-3 variants favors the notion that the Tom40 precursor appears in the 250-kDa form first and then progresses to the 100-kDa form. It has also been proposed that the Tom40 precursor in the 250-kDa intermediate is associated with the outer membrane on the intermembrane space side and is extractable with sodium carbonate (30). Our data from carbonate extraction of mitochondria following import of wild type Tom40 into wild type mitochondria do not allow conformation of this aspect of the model due to the unexpected finding that carbonate enhances assembly of Tom40 precursors in intermediates into the final 400-kDa form. However, the observation that precursors of the ⌬51-60 and ⌬321-3 variants of Tom40 in the 250-kDa intermediate were extractable by carbonate does support the notion that the Tom40 precursor is only peripherally associated with the membrane at this stage. One explanation for the action of sodium carbonate on wild type precursors might be that protein-protein interactions hold the molecule at an intermediate stage of assembly, and these are disrupted in the presence of carbonate. It is also conceivable that interactions between Tom40 molecules in existing TOM complexes FIG. 12. Wild type Tom40 precursor assembles into the TOM complex via an alternate pathway in mitochondria containing a mutant form of Tom40. A, radiolabeled wild type Tom40 precursor was imported for 20 min at either 0 or 25°C into mitochondria isolated from either control (74A) cells or cells expressing only the ⌬40 -48 mutant version of Tom40. Mitochondria were re-isolated and processed for BNGE. The gel was blotted to PVDF membrane and analyzed by autoradiography. B, samples were processed as in the legend for Fig. 9, except that radiolabeled wild type Tom40 precursor was imported into mitochondria isolated from cells expressing only the ⌬40 -48 mutant Tom40 protein.
Duplicates of the 0-and 10-min samples were subjected to sodium carbonate extraction following import and re-isolation of mitochondria (ϩcarb). C, samples were developed as in B, but after import and re-isolation, mitochondria were treated with proteinase K and processed for SDS-PAGE as described in the legend to Fig. 10. The positions and size of bands are indicated on the left of all panels. M, Tom40 monomer. may be weakened, facilitating replacement of existing subunits with incoming molecules.
Analysis of the import and assembly of wild type Tom40 precursors into mitochondria containing only the ⌬40 -48 mutant form of Tom40 revealed differences from the normal assembly pathway. Very little of the precursor was seen in the 250-kDa intermediate, and virtually none was detectable in the 100-kDa form. Instead, the precursor was found in a 450-kDa form, which appeared to assemble directly into the 400-kDa form. The simplest interpretation for the larger form is that it represents a precursor molecule associated with an existing TOM complex. Most of the precursor at this stage was extractable with sodium carbonate so that its initial association with the components in the high molecular weight form must precede integration into the membrane. Because there is no formation of the 100-kDa form, the precursor may integrate directly into the existing TOM complex, perhaps by displacing a pre-existing subunit. This interpretation suggests that amino acid residues 40 -48 of Tom40 molecules in the TOM complex may interact with incoming subunits. Although it may be surprising that an alternate pathway exists, it is also possible that a small amount of assembly occurs by this pathway under normal conditions but is not easily detected. Although the N terminus of Tom40 is not required for stable interactions between Tom40 molecules (56), our results suggest that the region is important for assembly of the protein into the TOM complex, because alterations in the N terminus of both Tom40 precursors and Tom40 molecules in the TOM complex have an affect on the assembly pathway.